Discovery and Optimization of Indoline-Based Compounds as Dual 5-LOX/sEH Inhibitors: In Vitro and In Vivo Anti-Inflammatory Characterization

The design of multitarget drugs represents a promising strategy in medicinal chemistry and seems particularly suitable for the discovery of anti-inflammatory drugs. Here, we describe the identification of an indoline-based compound inhibiting both 5-lipoxygenase (5-LOX) and soluble epoxide hydrolase (sEH). In silico analysis of an in-house library identified nine compounds as potential 5-LOX inhibitors. Enzymatic and cellular assays revealed the indoline derivative 43 as a notable 5-LOX inhibitor, guiding the design of new analogues. These compounds underwent extensive in vitro investigation revealing dual 5-LOX/sEH inhibitors, with 73 showing the most promising activity (IC50s of 0.41 ± 0.01 and 0.43 ± 0.10 μM for 5-LOX and sEH, respectively). When challenged in vivo in zymosan-induced peritonitis and experimental asthma in mice, compound 73 showed remarkable anti-inflammatory efficacy. These results pave the way for the rational design of 5-LOX/sEH dual inhibitors and for further investigation of their potential use as anti-inflammatory agents.


■ INTRODUCTION
Inflammation encompasses multiple physiological and pathological processes, triggered by converging pathways that generate a network of proinflammatory and proresolving mediators, involving different cell types and cellular responses. 1 The arachidonic acid (AA) cascade is a key biochemical pathway for pharmacologically targeting inflammatory diseases. Upon release from membrane phospholipids, AA is further metabolized via three divergent pathways: (a) transformation into proinflammatory prostaglandins (PGs) and thromboxane initially mediated by cyclooxygenase (COX) enzyme; (b) conversion into leukotrienes (LTs) mediated by 5-lipoxygenase (5-LOX) and by 5-LOX-activating protein (FLAP); and (c) metabolization by cytochrome P450 monooxygenases (CYP450) into the proinflammatory 20hydroxyeicosatetraenoic acid   2 and the antiinflammatory epoxyeicosatrienoic acids (EETs). The latter are hydrolyzed by soluble epoxide hydrolase (sEH) 2,3 to dihydroxyeicosatrienoic acids (DiHETrEs) with primarily proinflammatory effects. 4 The balance among these mediators determines the evolution of the inflammatory process toward progression and/or chronicity or toward resolution.
Traditionally, COX inhibitors (nonsteroidal anti-inflammatory drugs, NSAIDs) represent the reference drugs in this field and are also among the most used therapeutics worldwide. However, severe gastrointestinal or cardiovascular side effects upon prolonged use of these drugs may limit their therapeutic benefits. 5,6 Therefore, there are continuous efforts in the search for alternative pharmacological approaches characterized by a lower incidence of adverse reactions and improved efficacy. 7,8 LTs are proinflammatory mediators involved in many allergic 9−11 and nonallergic diseases. 12−15 Two different approaches have been adopted to reduce the LT-mediated response: the antagonism of the Cys-LT 1 receptor and the inhibition of LT biosynthesis. 16 Zafirlukast and montelukast, for instance, are Cys-LT 1 receptor antagonists approved for the treatment of asthma, rhinitis, and other allergic diseases.
However, these therapeutics are penalized by several side effects and by unsatisfactory biological responses in certain groups of patients. 17,18 On the other hand, the inhibition of 5-LOX or FLAP is considered a suitable strategy for decreasing the LT-mediated inflammatory response. The 5-LOX inhibitor zileuton is the only LT biosynthesis inhibitor, approved for the treatment of bronchial asthma, but due to its poor pharmacokinetics, its use is strongly limited by liver toxicity. 19 The third pathway of AA leads to anti-inflammatory EETs, which contribute to the homeostatic equilibrium of biological processes. In particular, they exert anti-inflammatory, 20 analgesic, 21 fibrinolytic, 22 antimigratory, 23 and angiogenic 24 effects. The tissue and plasma levels of EETs are increased by sEH inhibitors, which also block the DiHETrE synthesis, thereby counteracting inflammation. 25 Indeed, the use of sEH inhibitors is proposed as a proper therapeutic strategy in several pathological animal models. 25,26 Nevertheless, the clinical use of molecules interfering with one of the AA metabolic pathways is penalized by the wide pool of activities exerted by the downstream lipid mediators, which act as both proinflammatory and proresolving agents. 27 Furthermore, targeting a single enzyme in the eicosanoid cascade can cause substrate shunting and amplification of alternative pathways, resulting in decreased efficacy and increased side effects. 28 This is why polypharmacological approaches appear particularly useful in the design of anti-inflammatory drugs. 29,30 Multitarget drugs, indeed, accomplish a balanced modulation of eicosanoid levels and largely suppress shunting and/or redirection phenomena. 31−33 In this context, 5-LOX/sEH dual inhibitors offer the advantage of blocking the proinflammatory LT production and simultaneously increasing the antiinflammatory eicosanoid (i.e., EETs) levels. 34 In the present manuscript, we describe the design, the synthesis, and the pharmacological characterization of indoline-based dual 5-LOX/sEH inhibitors. The development of these molecules was carried out by an integrated approach of in silico and in vitro assays starting from an in-house molecular library. First, 53 bicyclic compounds from our library were cherry-picked and evaluated in silico, since the benzo[b]thiophene moiety is considered the pharmacophore of zileuton. 35 This approximation is due to the absence of zileuton/5-LOX experimentally resolved structure in the protein data bank (http:// www.rcsb.org/), 36 so only the binding mode, derived from recent molecular docking studies, could be taken into account. 37,38 From computer-aided analysis, nine derivatives were filtered and tested in vitro for their 5-LOX inhibitory properties leading to the identification of the indoline compound 43 as the hit compound. Therefore, we decided to design a second library of compound 43 analogues. Extensive in vitro testing of this library highlighted a remarkable dual inhibitory profile over 5-LOX and sEH for the synthesized molecules, with 73 emerging as the most potent compound. In silico studies also highlighted structure− activity relationship clues concerning the 5-LOX/sEH dual inhibition. When challenged in vivo, compound 73 showed remarkable efficacy in both peritonitis and asthma murine models, confirming the potential of these dual inhibitors in treating inflammatory diseases.
Derivatives 21, 43−46, and 48−52 were obtained, as shown in Scheme 3. Starting from 5-nitroindoline, different synthetic approaches were used to decorate both the N-1 and the C-5 positions. Using triphosgene and 2,2-dimethylpropan-1-amine, the urea derivative 19 was obtained in a 66% yield. Then, a continuous flow hydrogenation reaction provided the corresponding amine intermediate 20 (61% yield), which was converted to the final compound 21 (58% yield) through a reductive amination reaction with 4-fluorobenzaldehyde. Under the same conditions, N-1 alkylation of 5-nitroindoline was attained, using commercially available aldehydes, leading to intermediates 22−25 in 62−92% yields. Intermediates 26− 28 were synthesized using the same protocol and the modified aldehydes 26a, 27a, and 28a obtained, as described in Scheme 3. Continuous flow hydrogenation of these compounds afforded the corresponding amines 29−35 in 58−95% yields. Intermediates 29−35 were converted to isothiocyanates 36− 42 by reaction with CS 2 in toluene followed by treatment with ethyl chloroformate (48−77% yields). Reaction of these intermediates with 2,2-dimethylpropan-1-amine yielded thiourea analogues 43−49 in 52−67% yields. Compound 43 was further modified to guanidine 50 by reaction with HgO, in the presence of Na 2 SO 4 and CaCl 2 , followed by the addition of NH 4 OH, as previously described. 41 Compound 47 underwent Boc removal as described before, giving the final compound 51. The final compound 52 was obtained from 49, after removal of the MOM protecting group in acid conditions. Indoline derivatives 53−58 were synthesized according to Scheme 4. Intermediate 29 was converted to its carbamic chloride by reaction with triphosgene, which, upon reaction with cyclohexylamine and 2,2-dimethyl-1-propanamine, gave the final urea compounds 53 and 54 in 72 and 78% yields, respectively. Coupling of 29 with cyclohexanesulfonyl chloride afforded, instead, the final compound 55, as previously described. 41 The previously obtained intermediate 36 was converted to aromatic thiourea analogues 56−58 by the same method described above in 58−67% yields. 5-Nitroindoline was also decorated at the N-1 by acylation with 4-fluorobenzoyl chloride and 4-fluorophenylacetyl chloride to give intermediates 59 and 60 in 91 and 88% yields, respectively, and using di-tert-butyl dicarbonate to give the N-Boc intermediate 61 in an 88% yield (Scheme 5). The use of a continuous flow hydrogenation protocol followed by conversion to isothiocyanates 65−67 (58−62% yields) and reaction with 2,2-dimethylpropan-1-amine gave the final compounds 68 and 69, which were isolated in 65 The values are given as the mean ± standard error of the mean (SEM) of single determinations obtained in 3−4 independent experiments.
b Zileuton used as a positive control at 3 μM gives residual activities of 3.90 ± 4.14 and 14.24 ± 5.88% over PMNL and isolated 5-LOX, respectively.
Finally, the procedures used for the synthesis of final compounds 77 and 82 are described in Scheme 6. These molecules were synthesized to expand the structure−activity relationship (SAR) clues about dual 5-LOX/sEH inhibitors by increasing the planarity and aromaticity of the scaffold with the indole ring and allowing further exploration of the binding site by sterically hindered carbazole moiety. 5-Nitroindole was modified at N-1 using 4-fluorobenzyl bromide by the synthetic strategy previously described. 42 The corresponding intermediate 74 was subjected to the same sequential reaction steps described above to give the final thiourea compound 77 (34% overall yield).
Molecular Docking and Evaluation of 5-LOX and sEH Inhibition. Our in-house library (53 molecules) was in silicoscreened and based on visual inspection, energy score, and distance from catalytic iron ion (<4 Å); nine compounds were narrowed as potential inhibitors of 5-LOX. The bicyclic-based molecules, selected for the in silico preliminary screening toward 5-LOX, were structurally featured with (R)-dihydro-3H,5H-imidazo[1,5-c]thiazole-5,7(6H)-dione, 5-methylindolin-2-one, and indoline rings. For the spiro compounds, both enantiomers were considered in the calculations. The docking outcomes suggested that the indoline-based (43 and 55 and (R)-dihydro-3H,5H-imidazo[1,5-c]thiazole-5,7(6H)-dione 17 and 18) scaffolds are suitable molecular seeds for 5-LOX inhibitor design thanks to their ability to properly accommodate into the 5-LOX binding pocket (Table S1). The spiro compounds (4, 5, and 13) were the less promising scaffolds in silico (Table S1). For the sake of consistency, all derivatives were tested for their 5-LOX-inhibitory activity. At first, we evaluated their effectiveness in activated human polymorphonuclear leukocytes (PMNLs), using cell-based assays, which allows the analysis of the interference of the test compounds with 5-LOX in a biological environment. Results obtained were corroborated by testing compounds against the isolated human recombinant 5-LOX, which allows the identification of direct interference of the test compound with the target enzyme. The results are summarized in Table 1. Only derivative 43 reduced 5-LOX product levels in activated human PMNL (IC 50 1.38 ± 0.23 μM). In addition, 43 showed remarkable inhibitory activity against isolated 5-LOX (IC 50 0.45 ± 0.11).
The biological activity of 43 was in agreement with structural observations from molecular docking. In particular, it was observed that the indoline moiety is placed near the iron, hampering the access to the open position of the ion coordination sphere and also contributing to the complex lineup by van der Waals interactions with H372, H367, L368, L414, I415, F421, and L607. The thiourea group of 43 donates two H-bonds to the side chain of Q363 and accepts a hydrogen bond from Y181 ( Figure 1A). The neopentyl group of 43 establishes van der Waals contacts with Y181, F421, A424, N425, P569, H600, and A603. The 4-fluorobenzyl moiety is engaged in an aromatic H-bond with the N407 side chain and van der Waals contacts with W147, F151, H372, L368, L373, A410, R411, and I415. Furthermore, the fluorine interacts with the side chain of R411. The remarkable 5-LOX inhibitory activity of 43 together with the observed intermolecular interactions led to the design of 19 structurally correlated compounds.
Specifically, we structurally modified the tert-butyl moiety (56−58), the thiourea group (21,50,53,54), the indoline at C-2 and C-3 (77, 82), and the 4-fluorobenzyl group (44−49, 51, 52, 68, 69, 73) to get clues about structure−activity relationships. All compounds were tested by different in vitro assays. Initially, their capability of reducing 5-LOX activity both in activated PMNL and of isolated 5-LOX was challenged. Then, considering the importance of the sEH enzyme in the AA cascade, we questioned whether the parental molecule (compound 43) and its analogues could affect the activity of this enzyme. We started evaluating 43/sEH in silico interaction, observing a good fit into the binding cavity of sEH by compound 43 (Figure 2A). The docked pose of 43 shown gives π−π interactions with H524 and W525 by indoline and 4-fluorobenzyl moieties, respectively. Its urea group donates a hydrogen bond to D335, whereas the neopentyl moiety gives van der Waals contacts with W336, M339, Q384, Y466, and L499.
The docking hypothesis was confirmed by in vitro results that revealed an IC 50 of 1.39 ± 0.45 μM against the human isolated sEH (Table 2).
Considering this data, all of the indoline compounds were tested for their inhibitory activity on isolated sEH. Results obtained are summarized in Table 2.
Compounds 53, 54, and 73 exhibit the best experimental outcomes, filling equivalent spaces at the 5-LOX binding site when compared to 43, and display the same pattern of intermolecular contacts by the common structural moieties ( Figure 1). In contrast to the parent compound 43, in 53 and 54 the thiourea is substituted with urea that preserves the same network of H-bonds in the 5-LOX binding pocket. Like the neopentyl group of 43, the cyclohexyl substituent of 53 gives van der Waals contacts with Y181, F421, A424, N425, P569, H600, and A603. Compounds 53 and 54 inhibited the isolated 5-LOX enzyme potently, with an IC 50 of 0.28 ± 0.02 and 0.18 ± 0.05 μM, respectively ( Table 2). With respect to 43, both compounds showed no 5-LOX inhibitory activity in the cellbased assay. This might be explained by poor membrane permeation, but different reasons cannot be excluded. Moreover, 53 and 54 also showed a potent inhibitory effect against sEH, with an IC 50 of 61 ± 3 and 100 ± 10 nM, respectively ( Table 2). This data is in accordance with the literature suggesting that the presence of an urea group is a pivotal requisite for potent sEH inhibition. 43 Indeed, the two compounds fitted very well in the binding cavity of the enzyme, presenting superimposable conformations with the cocrystallized ligand 34N ( Figure 2B,C). 44 For both binders, the indoline and 4-fluorobenzyl moieties give π−π interaction with H524 and W525, respectively. Their urea group is involved in a network of four hydrogen bonds with the side chains of D335, Y383, and Y466, unlike 43 that donates a hydrogen bond with its thiourea group ( Figure 2A). The neopentyl (54) and cyclohexyl (53) groups are involved in van der Waals contacts with Y336, M339, Q384, Y466, and Leu499. It is noteworthy that the contacts found for the indoline, urea, and neopentyl/cyclohexyl are also observed with 34N. 44 In 50, the thiourea was replaced by a guanidine, losing the hydrogen bond with the Y181 side chain. The loss of interaction with Y181 seems responsible for a partial loss of activity for compound 50 showing three times higher IC 50 on isolated 5-LOX compared to 43 (1.42 ± 0.23 vs 0.45 ± 0.11 μM).
The most interesting compound of the series is represented by compound 73. It is noteworthy that 73 structurally differs from all other compounds for the NO group instead of fluorine ( Figure 1D). The NO group is engaged in hydrogen bonds with side chains of R411, tightening the affinity toward 5-LOX. The phenyl ring in the two aromatic H-bonds with the backbone CO of L368 and N407 shows the highest potency against both isolated 5-LOX and PMNL (0.41 ± 0.01 and 0.59 ± 0.09 μM, respectively; Figure 2D and Table 2). Moreover, 73 also exhibited an important efficacy against sEH with an IC 50 of 0.43 μM ± 0.10 ( Table 2). We observed that the NO All values are given as the mean ± SEM of single determinations obtained in 3−4 independent experiments. b Zileuton used as a positive control at 3 μM gives residual activities of 3.90 ± 4.14 and 14.24 ± 5.88% over PMNL and isolated 5-LOX, respectively. AUDA used as a positive control at 1 μM gives a residual activity of 3.90 ± 4.14% over isolated sEH. c nd, not determined.
group also favors the binding against sEH by establishing two H-bonds with the backbone NH group of F497 and H524. With respect to 73, the phenol moiety of 52 is only Hbonded to the backbone CO of N407, justifying a reduced activity on the isolated 5-LOX and PMNL (IC 50 = 5.10 ± 2.92 and 2.90 ± 0.75 μM, respectively); otherwise, sEH inhibition is maintained with an IC 50 of 0.79 ± 0.52 μM (Table 2). In fact, even though the NO and OH groups of 73 and 52 are also hydrogen-bonded to the backbone NH and CO of H524 and V416 in the sEH binding cavity, their presence displaces the thiourea moiety, compared to urea of 53 and 54, giving rise to only one H-bond (Figure 2), justifying the mild reduction of sEH activity by 53 and 54.
If compared with 43, compounds 44−49 and 51 present a substantial modification of the substituent at indoline nitrogen that could affect the correct binding into the 5-LOX catalytic site. Specifically, the switch from methylene (43) to ethyl linker between 4-fluorophenyl and indoline moieties (44) is well tolerated, while a further increase of the substituent size (45, 46, and 49) impairs the interaction from the remaining structural portions of the small molecules, especially for 49. Compound 48 showed a superimposable accommodation of common molecular portions with the parent compound in the 5-LOX binding site, but the hydroxybutyl chain is quite folded although H-bonded to the backbone CO of Q363. All of these compounds exhibit a micromolar activity against the isolated 5-LOX, except 44, which showed an IC 50 comparable to 43 (0.38 ± 0.05 vs 0.45 ± 0.11 μM). However, for derivatives 45 and 48, a decrease of effectiveness against sEH was observed, while their analogues 44, 46, and 49 act on the same target with an IC 50 in the low micromolar range. Unfortunately, none of them are effective against 5-LOX in PMNL. Compound 49 binding to sEH could be rationalized using the same considerations made for 53 and 54. The compound accepts a H-bond from the backbone NH of H524, leading to an unfavorable entropic loss ( Figure 2). Like 52 and 73, increasing the size of the substituent at indoline nitrogen (44 and 46) also allows hydrogen bonding with only the side chain of D335 ( Figure 2). Derivative 45, endowed with a larger substituent at indoline nitrogen than 44 and 46, is unable to give hydrogen bonds by the thiourea moiety ( Figure 2).
The conversion of the indoline core to indole (77) induces a binding conformation in 5-LOX characterized by less contacts with W147, F151, H372, L368, L373, A410, R411, and I415, reflecting an almost complete loss of 5-LOX and sEH inhibition. In 82, the indoline was converted into the more hindered carbazole, resulting in docked poses with distorted structural moieties, especially for thiourea and tricyclic aromatic portion, inducing a total inactivity against both targets. For 56−58, a π−π interaction is observed among the phenyl ring and side chain of F421, unlike the parent compound. Moreover, 56 gives a π−π interaction with F359, while 57 and 58 give a π−π interaction with Y181. The compounds maintain good activity in the cell-free assay, but only 56 and 57 are able to affect the 5-LOX activity in PMNL, likely due to their higher lipophilicity than the acid analogue 58. In consideration of the data described, taking particularly into account the activity of the synthesized derivatives on 5-LOX in PMNL, more properly resembling the biological environment, we have decided to select compound 73 for Values represent means ± SEM; n = 6 mice for each group. Data were analyzed by one-way analysis of variance (ANOVA) and Bonferroni. Statistical significance is reported as follows:°°P < 0.01 and°°°P < 0.001 vs control; * P < 0.05; ** P < 0.01; and *** P < 0.001 vs zym + vehicle.
further pharmacological characterization of this class of compounds.
Evaluation of COX-1 and COX-2 Inhibition in Intact Cells. We next investigated the impact of 73 on COX-1 and COX-2, enzymes within the AA cascade that are involved in the biosynthesis of prostanoids in addition to 5-LOX and sEH. A well-established in vitro cell culture assay (J774 murine macrophages) was performed to evaluate the effects of 73 against both COX isoforms. 45−47 Stimulation of J774 macrophages with AA (15 μM) for 30 min induced a significant increase of PGE 2 levels in comparison to unstimulated control cells. 73 weakly inhibited the production of PGE 2 primarily generated via COX-1 at micromolar concentrations ( Figure  S76A). The same trend was observed for PGE 2 production in LPS-stimulated cells in the absence ( Figure S76B) or presence ( Figure S76C) of AA, which is mainly mediated by inducible COX-2. Indomethacin and celecoxib as reference agents were active as expected. In addition, 73 did not affect COX-2 expression induced by the stimulation of cells with LPS (10 μg/mL) ( Figure S76D). These results indicate a higher target selectivity of compound 73 for 5-LOX and sEH enzymes.
Finally, cytotoxic effects were excluded since 73 did not impair cell viability at all tested concentrations ( Figure S76E). Overall, molecular docking and in vitro biological investigations suggest 73 as a promising drug candidate for further in vivo pharmacological studies.

Evaluation of In Vivo Anti-Inflammatory Effects. Compound 73 Reduces Inflammation in Zymosan-Induced
Peritonitis. The anti-inflammatory efficacy of 73 was evaluated in vivo in zymosan-induced mouse peritonitis, an experimental model of acute inflammation related to LTs and other lipid mediators. 48,49 Zileuton and AUDA were used as controls (i.p. 10 mg/kg, 30 min before zymosan; Figure 3A). During the onset of inflammation, zymosan activates resident murine peritoneal macrophages that produce LTC 4 . The progressive phase of inflammation is instead dominated by infiltrated neutrophils, which generate the potent chemoattractant LTB 4 and other proinflammatory mediators such as PGE 2 , nitric oxide, and TNF-α. Accordingly, 30 min and 4 h after zymosan injection, a significant increase of LTC 4 and LTB 4 was observed as compared to the unstimulated control group ( Figure 3B,C). The i.p. pretreatment of mice with 73 (10 mg/ kg, 30 min before zymosan; Figure 3A) significantly reduced LTC 4 and LTB 4 levels in the peritoneal exudate, comparable to zileuton ( Figure 3B,C). Since LTB 4 is a major chemoattractant for leukocytes, 73 caused a concomitant reduction of leukocyte recruitment in the peritoneal cavity ( Figure 3D). Surprisingly, in compound 73, in pretreated animals, a strong reduction of PGE 2 levels was observed in comparison to vehicle-treated mice ( Figure 3E), apparently in contrast with in vitro data ( Figure 76S). Nevertheless, a closer look at the in vitro data reveals that compound 73 is able to exert a small but significant reduction of PGE 2 at the highest concentration used (10 μM; Figure 76S). This is why the high local concentration reached in the peritoneum immediately after intraperitoneal administration could be accounted for the inhibitory effect over PGE 2 levels. In addition, compound 73 also showed in vivo anti-inflammatory effects by the inhibition of zymosan-induced NOx ( Figure 3F) and TNF-α ( Figure 3G) in the peritoneal exudates of zymosan-treated mice.
Since compound 73 contains a nitrosobenzene and a dihydroindole moiety, the overall stability of the molecule can be questioned. This is why we decided to challenge the compound stability in mouse plasma. To this aim, blood samples were collected from animals at predetermined intervals (0.5, 1, 2, and 4 h) and, upon extraction, were analyzed by high-performance liquid chromatography/mass spectrometry (HPLC/MS) to quantify compound 73. Results obtained (Table S3) show substantial stability of compound 73, with a plasma half-life in mice of 1.6 h.
Compound 73 Relieves Hallmarks of Asthma and LT Pulmonary Levels. Since LTs play a pivotal role in the pathogenesis of asthma by inducing immune cell infiltration, pulmonary inflammation, and bronchoconstriction, 15 we investigated the effects of 73 in an experimental model of asthma. Mice were pretreated with 73 i.p. 30 min before ovalbumin (OVA) injection on days 0 and 7. Animals were sacrificed after 21 days to evaluate bronchial hyper-reactivity, pulmonary inflammation, pulmonary LTC 4 levels, plasma IgE, Values represent means ± SEM; n = 6 mice for each group. Data were analyzed by two-way ANOVA plus Bonferroni (B and C) and one-way ANOVA plus Bonferroni (E−I). Statistical significance is reported as follows:°P < 0.05;°°P < 0.01; and°°°P < 0.001 vs control; ** P < 0.01 and *** P < 0.001 vs OVA + vehicle. Black arrow, bronchial epithelium thickness; red arrow, pulmonary cell infiltration in peribronchial areas; and asterisk; pulmonary cell infiltration in perivascular areas (D). and Th2 cytokine production ( Figure 5A). OVA sensitization induced airway hyper-reactivity to carbachol ( Figure 5B), and increased bronchial relaxation in response to salbutamol was observed ( Figure 5C). Intraperitoneal treatment of mice with 73 reversed OVA-induced bronchial hyper-reactivity to carbachol ( Figure 5B) and fully restored the adrenergic bronchial relaxation induced by salbutamol ( Figure 5C). OVA sensitization caused airway inflammation by inducing a morphological alteration ( Figure 5D) and increasing the bronchial epithelium thickness (black arrow) ( Figure 5D,E). Further, OVA sensitization promoted pulmonary cell infiltration in peribronchial (red arrow) and perivascular areas (asterisk) compared to the control group ( Figure 5D). Pretreatment with 73 significantly reduced the epithelial thickness ( Figure 5E) in OVA-sensitized mice. The beneficial effect of 73 on lung function was associated with the reduction of pulmonary LTC 4 levels in sensitized mice treated with 73 ( Figure 5F). However, 73 did not affect sensitization mechanisms. Indeed, 73 did not modulate plasma IgE levels ( Figure 5G) and pulmonary T-helper type 2 cytokines such as interleukin-13 and interleukin-4 ( Figure 5H,I) in OVAsensitized mice. In the same setting of experiments, the effects of 73 were compared to those of zileuton ( Figure 5). The data obtained provide similar efficacy of the two molecules; however, the required dose of zileuton is 35 mg/kg, 50 while 10 mg/Kg of compound 73 generate a pharmacological response. The improved efficacy in terms of pharmacological activity reflects the synergic effect obtained by interfering with the two enzymes than a single target.
General Procedure A: Thiazolidinone Synthesis (1−3). 5-Methylisatin (1.0 mmol) was dissolved in ethanol (50 mL), and the solution was warmed at 100°C. Mercaptoacetic acid (1.5 mmol) and 4-(aminomethyl)benzoic acid, or 4-aminobutanoic acid, or N-Boc-1,3-propanediamine (0.5 mmol) were added, and the mixture was stirred for 120 min. Then, 5.0 mL of a solution of NaHCO 3 (10% v-v) was added and the organic phase was evaporated in vacuo. The crude was dissolved in dichloromethane, and a basic aqueous solution (Na 2 CO 3 2 N) was employed to wash the organic phase (3 × 100 mL). The dichloromethane layer was then dried on Na 2 SO 4 , filtered, and evaporated under vacuo. Flash chromatography on silica gel using different eluent systems yielded intermediates 1−3.

Journal of Medicinal Chemistry pubs.acs.org/jmc Article
To this solution, 0.23 mmol of NaH were added portionwise and the mixture was allowed to react for 30 min. Then, 0.23 mmol of methyl iodide or 4-fluorobenzyl chloride in DMF were added dropwise and the reaction was warmed to room temperature and maintained under stirring for a further 12 h. Then, the reaction was quenched with 10% aqueous solution of citric acid and washed with brine. The organic layer was separated, dried over anhydrous Na 2 SO 4 , filtered, and evaporated in vacuo. The crude product was purified by flash chromatography using n-hexane/ethyl acetate (4:1 v-v) as the mobile phase to obtain intermediates 11, 74, and 78. General Procedure D: Boc Removal (7, 10, 17, 18, 51, 71). The N-Boc-protected intermediates (0.2 mmol) were dissolved in a mixture of TFA/DCM (1/3, v/v), and triisopropylsilane (TIS, 0.05 mmol) was added. The reaction was stirred at room temperature for 2 h. Then, a solution of NaOH (2 N) was added until pH 7. The mixture was diluted with water and dichloromethane, and the organic phase was extracted, dried over Na 2 SO 4 , filtered, and concentrated under vacuum. The intermediates obtained were not further purified.
General Procedure F: Urea Formation (19,53,54). Aminic compounds (5-NO 2 -indoline or 29, 0.1 mmol) were dissolved in dichloromethane, and triphosgene (0.025 mmol) and triethylamine (0.12 mmol) were added. The mixture was reacted for 30 min, and the second amine (2,2-dimethylpropan-1-amine or cyclohexylamine) was introduced and the reaction was stirred for 1 h at room temperature. Then, the organic solvent was treated with water (3 × 100 mL) and the organic phase was dried over Na 2 SO 4 , filtered, and evaporated. Ureidic compounds were isolated after flash chromatography using different ratios of n-hexane/ethyl acetate as the mobile phase.
General Procedure G: Reductive Amination (21, 22−28, 72). 5-NO 2 -indoline derivative (0.1 mmol) was dissolved in a solution of DCM/CH 3 COOH (5:1 v/v) at room temperature. To this solution, an amount of 0.2 mmol of proper aldehyde were added and the mixture was warmed to reflux for 1.5 h. Then, an amount of 0.18 mmol of sodium triacetoxyborohydride were added portionwise and the mixture was allowed to reflux for a further 3−5 h. After cooling to room temperature, NaOH(1N) was added. The organic phase was separated and extracted one more time with the alkaline solution. Then, it was dried over Na 2 SO 4 , filtered, and concentrated in vacuo. The crude products were purified by column chromatography using different mixtures of n-hexane/ethyl acetate as eluent.
General Procedure H: Continuous Flow Hydrogenation (20, 29−  35, 62−64, 75, 80). Reduction of 5-nitroindoline, 5-nitroindole, and 5-nitrocarbazole derivatives was performed by continuous flow hydrogenation employing the H-Cube hydrogenator and commercially available Pd/C 10% cartridges as a catalyst. Initial nitro compounds were dissolved in a mixture of tetrahydrofuran (THF)/ CH 3 OH (1:1, v/v) at a final concentration of 0.1 M and were pumped at a flow rate of 1.0 mL/min. The temperature was set at 30°C , while the hydrogen inlet pressure was set at 10 bar. Finally, the reaction solution was evaporated in vacuo and the obtained products were used in the following step without further purification.
General Procedure I: Synthesis of Isothiocyanide (36−42, 65−  67, 76, 81). The proper amine (0.1 mmol) was dissolved in toluene, and 0.1 mmol of triethylamine and 0.2 mmol of carbon disulfide were added and reacted overnight. Subsequently, the organic phase was concentrated in vacuo and the crude was dissolved in dichloromethane, and 0.1 mmol of triethylamine (TEA) and 0.1 mmol of ethyl chloroformate were added and the mixture was stirred for 12 h at room temperature. Then, an aqueous solution (10% w/w) of NaHCO 3 was added (3 × 100 mL) and the extracted organic solvent was dried on Na 2 SO 4 , filtered, and evaporated. Isothiocyanide derivatives were obtained after flash chromatography using nhexane/ethyl acetate as eluent.

1-(1-(4-Fluorobenzyl)indolin-5-yl)-3-neopentylurea
For binding investigation toward 5-LOX, the Induced Fit Docking 57−59 was applied, using the extended protocol, generating 80 ligand−protein poses at XP precision. The grid was centered on Fe 2+ with an inner box of 10 Å, and the outer one was automatically generated. The conformational search was performed allowing the sample ring conformations of the small molecules, with an energy window of 2.5 kcal/mol. For Prime refinement, the default values were used. For docking calculations on sEH, Glide software 60,61 was employed to dock the ligand against sEH. To validate the docking methodology, the cocrystallized ligand 34N with sEH was docked and the obtained conformation was compared with the experimental one (RMSD = 0.486 Å). 62−64 The inner and outer receptor grid boxes of 10 Å and 17 Å, respectively, centered on the x, y, and z coordinates: 75.42, −9.30, and 68.12. In the first step, Standard Precision (SP) was applied along with default parameters, producing one pose per ligand. These poses from the SP calculations were utilized as input conformations for three rounds of predictions in the Extra Precision (XP) Glide mode: flexible ligand; only amide bond trans conformation allowed; nitrogen inversion and ring conformations (with an energy cutoff of 2.5 kcal/mol) sampling. The enhanced sampling mode was utilized, saving 10,000 poses/ligand for the initial docking step and 1000 poses/ligand for energy minimization. One thousand maximum output structures/ligands were kept applying 0.8 as the scaling factor for van der Waals radii and 0.15 as the partial charge cutoff. Postdocking optimization was executed on docked poses, filtering through 100 maximum number of poses and 0.5 kcal/ mol cutoff to reject the obtained minimized conformations. The energy contributions of the Epik state penalty, aromatic bonds, and intramolecular H-bond reward were considered in the predictions. The docking outcome analysis and figure preparation were carried out by Maestro (version 11).
Cell-Free 5-LOX Activity Assay. Human recombinant 5-LOX Western Blot Analysis. The analysis of COX-2 in J774 macrophages was performed on whole cell lysates. After stimulation with LPS for 24 h, cells were washed with cold PBS, collected by scraping, and centrifuged at 8000 rpm for 5 min at 4°C. Pellets were lysed by syringing with RIPA buffer (Trizma Base, NaCl, 100 mM EDTA, 10% Na-deoxycholate, and 10% Nonidet P-40), completed with 200 mM of activated orthovanadate and complete protease inhibitor cocktail (Sigma-Aldrich), and centrifuged at 12,000 rpm for 10 min at 4°C. The supernatants were collected, and protein concentration in cell lysates was determined by Bio-Rad Protein Assay (Bio-Rad). Equal amounts of protein (50 μg) were mixed with gel loading buffer (50 mM tris, 10% sodium dodecyl sulfate (SDS), 10% glycerol, 10% 2-mercaptoethanol, and 2 mg/mL of bromophenol) in a ratio of 4:1, boiled for 5 min. Each sample was loaded and electrophoresed on a 10% SDS−polyacrylamide gel. The proteins were transferred onto nitrocellulose membranes (0.2 μm nitrocellulose membrane, Trans-Blot TurboTM, Transfer Pack, Bio-Rad Laboratories). The membranes were blocked with 0.1% PBS-Tween containing 5% nonfat dry milk. After blocking, the membranes were incubated with the relative primary antibody overnight at 4°C. Mouse monoclonal antibody anti-COX-2 (BD Transduction Laboratories) was diluted to 1:1000 in 0.1% PBS-Tween, 5% nonfat dry milk; mouse monoclonal antibody anti-β-actin (Santa Cruz Biotechnology) was diluted to 1:1000 in 0.1% PBS-Tween, 5% nonfat dry milk. After the incubation, the membranes were washed three times with 0.1% PBS-Tween and were incubated for 2 h at room temperature with horseradish peroxidase-conjugated antimouse secondary antibody (Santa Cruz Biotechnology) diluted to 1:2000 in 0.1% PBS-Tween containing 5% nonfat dry milk. The membranes were washed, and protein bands were detected by an enhanced chemiluminescence system (ChemiDoc, Bio-Rad). Densitometric analysis was performed with Image Lab software (Bio-Rad Laboratories).
Cell Viability. Cell respiration, an indicator of cell viability, was assessed by the mitochondrial-dependent reduction of 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich, Milan, Italy) to formazan. Cells were plated to a seeding density of 1.0 × 10 5 in 96-multiwell. After stimulation with test compound for 24 h, cells were incubated in 96-well plates with MTT (0.2 mg/mL) for 1 h. Culture medium was removed by aspiration, and the cells were lysed in DMSO (0.1 mL). The extent of reduction of MTT to formazan within cells was quantified by the measurement of OD 550 .
Animals. Male CD-1 mice (33−39 g, 8 weeks, Charles River Laboratories; Calco, Italy) and female BALB/c mice (20 g, 8 weeks, Charles River Laboratories) were fed with standard rodent chow and water and acclimated for 4 days at a 12 h light and 12 h dark schedule in a constant air-conditioned environment (21 ± 2°C). Mice were randomly assigned to groups, and experiments were carried out during the light phase. Experimental procedures were conducted in conformity with Italian (D.L. 26/2014) and European (directive 2010/63/EU) regulations on the protection of animals used for scientific purposes and approved by the Italian Ministry.
In Vivo Plasma Levels of 73. For the analysis of plasma levels of 73, mice (5 per group) received 10 mg/kg of i.p. injection in a volume of 500 μl of 2% DMSO in saline. After selected time points, mice were sacrificed (CO 2 atmosphere) and blood (approximately 0.7−0.9 mL) was collected by intracardiac puncture using citrate as an anticoagulant. Then, plasma was obtained by centrifugation at 800g at 4°C for 10 min and immediately frozen at −80°C. For the extraction of 73 and the internal standard (IS) tolbutamide from mouse plasma, the method of protein precipitation was employed. The frozen mouse plasma samples were thawed at room temperature. Ten microliters of IS solution (2 μg mL −1 ) were added at 100 μL of plasma and vortexed for 30 s. Then, 890 μL of ice-cold methanol was added to all tubes and the extraction was performed by vortex mixing for 5 min, followed by centrifugation for 10 min at 14,600 rpm at 25°C . The supernatants were collected, filtered by 0.45 μM of RCmembranes, and then injected in UHPLC-MS/MS. Stock solutions of 73 and tolbutamide were prepared in DMSO (1 mg mL −1 ). The follows: 0.0−6.0 min, 1−55% B; 6.00−6.50 min, 55−95% B; 6.50− 7.50 min isocratic to 95% B; in 3.5 min returning to 1% B and remained unchanged at the end of the run. The total run time was 11.5 min.
Eicosanoid quantification was carried out in multiple reaction monitoring (MRM) mode monitoring transition from deprotonated precursor to product. For optimization of the mass spectrometer standards, 5,6-EET, 8,9-EET, 11,12-EET, 14,15-EET, 5,6-DHET, 8,9-DHET, 11,12-DHET, 14,15-DHET, 11,12-DHET d 11 , and 14,15-EET d 11 at a concentration of 500 ng mL −1 in EtOH were individually introduced into the mass spectrometer with flow injection mode. As the m/z 319 and m/z 337 precursor ions were obtained for all EETs and DHETs, respectively, the most specific ion product transition was used for quantification, as reported in Table S4. The dwell time was set to 100 ms for all of the monitored transitions. Interface temperature, DL temperature, and heat block temperature were set to 300, 250, and 350°C, respectively. Nebulizing gas, heating gas, and drying gas flows were set to 3, 10, and 10 L min −1 , respectively. All of the data was collected in the centroid mode and acquired and processed using Lab Solution workstation software.  (Tables S5 and S6) was performed using linear regression of the response ratios (peak area analyte/peak area internal standard) obtained from the calibration curve to calculate the corresponding eicosanoid amount. The extraction of eicosanoids from peritoneal exudates was performed, as previously reported. 69 Briefly, 500 μL of peritoneal exudates were diluted to 1 mL with phosphate salt buffer and spiked with ISs. The eicosanoids were extracted using Strata-X reversed-phase SPE columns (Phenomenex, Bologna, Italy). After loading the sample, the columns were washed with 10% MeOH, and the eicosanoids were then eluted with MeOH. Prior to the LC-MS/MS analysis, the eluent was dried under vacuum using a SpeedVac (Savant, Thermo Scientific, Milan, Italy) and dissolved in 50 μL of UHPLC solvent A.
Experimental Model of Murine Asthma. BALB/c mice were treated with 0.4 mL s.c. of a suspension containing 100 μg of ovalbumin (OVA) absorbed to 3.3 mg of aluminum hydroxide gel on days 0 and 7. 70,71 Compound 73 (10 mg/kg), zileuton (35 mg/kg), or vehicle (dimethyl sulfoxide 4%, 0.5 mL) was administered i.p. 30 min ( Figure 5A) before each OVA administration. Animals were sacrificed on day 21 by an overdose of enflurane, and lungs, bronchi, and plasma were collected. In particular, blood was collected by intracardiac puncture using citrate as an anticoagulant. Then, plasma was obtained by centrifugation at 800g at 4°C for 10 min and immediately frozen at −80°C. 72 Total IgE levels were measured by ELISA kit (BD Biosciences, Pharmingen, San Jose, CA). Bronchi were cut in rings of 1−2 mm in length, placed in organ baths, and fixed to an isometric force transducer 7006 connected to a Powerlab 800 (AD Instruments, Ugo Basile, Comerio, Italy). After stretching the rings to a resting tension of 0.5 g and equilibration for at least 30 min, the rings were challenged with carbachol (1 μM) until a reproducible response was observed. To assess bronchial reactivity, the cumulative response to carbachol (0.001 to 3.16 μM) and salbutamol (0.01−30 μM) was measured. 50 Results were expressed as dyne per mg tissue.
The right lung lobes harvested from mice were rapidly fixed in 4% formalin. The tissues were embedded in paraffin, and cryosections of 7 μm were cut. The slices were processed to remove the paraffin, and following rehydration, hematoxylin and eosin (H&E) staining was performed. Sections were analyzed using a Leica Microsystem with a scale bar of 50 μm (H&E). Pulmonary cell infiltration and epithelial thickness were evaluated using ImageJ Fiji software. 50 Left lungs were isolated and homogenized in PBS (Sigma-Aldrich, Milan, Italy). The homogenate was centrifuged (4°C, 6000g, 10 min). 50 The levels of LTC 4 (Cayman Chemical, BertinPharma, Montigny Le Bretonneux, France), IL-13, and IL-4 (Invitrogen, Vienna, Austria) were measured by ELISA according to the manufacturer's instructions. The levels of LTC 4 , IL-13, and IL-4 were expressed as pg/mL. Statistical Analysis. The results are expressed as the mean ± SEM of n observations, where n represents the number of animals or number of experiments performed on different days. Statistical evaluation was performed by one-way or two-way ANOVA using GraphPad InStat (Graphpad Software Inc., San Diego, CA) followed by a Bonferroni post hoc test for multiple comparisons, respectively. Post hoc tests were run only when F achieved P < 0.05, and there was no significant variance in the homogeneity. A P value <0.05 was used to define statistically significant differences between mean values. ■ ASSOCIATED CONTENT * sı Supporting Information